 Live from Las Vegas, it's theCUBE, covering HPE Discover 2017, brought to you by Hewlett Packard Enterprise. Welcome back, everyone. Live here in Las Vegas, this is theCUBE's special presentation of Hewlett Packard Enterprise, HPE Discover 2017. I'm John Furrier with theCUBE, my cohost, Dave Vellante for the next three days of wall-to-wall coverage here in Las Vegas. Our next guest is Tom Latin, Vice President and General Manager of HPE Server Options, which is all the good stuff that wrap around the servers. It's certainly the big news here at HPE is the Gen10, the continuation of the generation of servers, which is all the rage these days. People talking about servers, are we buying more or buying less? Still a lot of private cloud going on, nothing really changing in the premise world. Welcome to theCUBE. Thank you, good to be here. All right, so what's the big news? I mean, to me, the thing that I've always been watching is the Gen8 stuff, Gen10 now is the new announcement platform for servers. Certainly. A lot of stuff happening around, a lot of innovations. Give us a quick update on the key things happening around Gen10 and the options. Well, certainly with Gen10, we've got a whole new level of security that is a big part of the Gen10 story. And with the server options in the memory subsystem and the storage subsystem and the networking subsystem, all of those technologies contribute to building up those layers of security in Gen10. So that's a key part of what we're doing this week at Discover. The second thing in the options space, the server options space is persistent memory. So we introduced persistent memory on Gen9 servers last year with an NVIDIM, an eight gigabyte NVIDIM. We're extending that portfolio this year with the Gen10 servers and increasing the NVIDIM capacity to 16 gigabytes. But the really big news in the persistent memory offering is scalable persistent memory. I just saw a great article on CRN, computer reseller news or CRN, really giving you guys some props on this. Persistent memory combines the performance of DRAM with the persistence of traditional SSDs or spinning disk. So essentially a huge performance gain and it's always good. HP has that, always has some good mojo when it comes to high quality products and performance. But talk about the impact because one of the things that's going on right now is storage has to be invisible, but applications now have going beyond virtual machines. You've got containers, you've got all kinds of cloud things happening with the app. So the data, the state of the data and the performance of the data is really critical for customers. What's the impact of this to that trend? This is persistent memory unleashes an incredible amount of performance for the server, for the application running on the server. At its most fundamental level, persistent memory and NVIDIM or a scalable persistent memory implementation can replace a layer of flash storage. And we're seeing performance benefits on the order of doubling the performance of the application just by swapping out an SSD for a persistent memory or an NVIDIM. When the applications, when the architects actually modify the applications to take advantage of the fact that there's a persistent memory layer in there, we see performance benefits as small as four times improvement. But in many cases, we're seeing 20, 25, 27X performance improvements because the architects of the applications now can dive right in to the memory subsystem without going through layers and layers of code and moving data around to get to storage. Love the options concept you guys are running because I think one of the things that we hear is, you know, flexibility is key and it's kind of like going to the store and like getting some accessories for your suit. Not one size fits all. You need to have a lot of options given the environment. So I got to ask you the competitive question. How do you guys compare vis-a-vis the competition with the persistent memory stuff, for instance? You mentioned the performance. How does that compare with some of your competitors? I think persistent memory is a good example of where we're way out ahead of where our competition is right now. It's, you can't just drop in a piece of hardware in the server and get all of the benefit out of it. You've got to be able to integrate that piece of hardware with the system software, the BIOS in the system, and then work with software application partners, especially to go modify or to optimize their applications to work best with that component that's been in there. And you guys do that. We do that extensively. Not the customer, customer just, what? Customer gets it, it just works. So what's the typical sort of anatomy of a life cycle of a server these days? And I mean it used to be in the old days it was called peripherals and peripherals made up probably 60 to 70% of the market. So it was very huge, you know, quite a huge opportunity. So how does it work from a customer standpoint? Do they, what's their starting point? How do they plan this out? Or is it more reactive? I wonder if you could sort of bring us up to date on that dynamic, Tom. I think that's part of what we bring to the market is a set of server platforms that have an incredible breadth of capability and that breadth comes from the configuration choices that the customer can make through the server options portfolios. So generally there are a customer will standardize on a few different platform types and then deploy those servers for a variety of different workloads. And so it's through those configuration choices, storage capacity, memory capacity, the performance of different layers in the memory and storage hierarchy that allow them to be able to fine tune a small set of servers really for a much broader set of workloads. So we've commented for a number of years now on the cube that the pendulum is swinging. Storages and servers are coming and computer coming closer together. Absolutely. You certainly saw that with Flash and PCIe, sort of those trends brought storage and compute together. Now you're accelerating that even further with persistent memory. So as that happens, one of the big challenges is data sharing, right? Now you hear things like NVMe over fabric and other technologies to link these capabilities, these nodes, if you will. I mean, it sounds like the machine. So what's happening there? And help us sort of squint through those two big trends. Yeah, I think that as we evolve architectures to focus on the data and recognize that the data really is where the value resides, we're moving from a world where, first of all, sharing the data on storage makes a lot of sense to enable that. If you look at some of the demonstrations that we've done recently with the machine and with memory-driven computing, it's about taking that shared presence of the data or shared instance of the data to an entirely new level where the processing capability can get directly at it and operate it and work on that really as a shared body of information, shared body of knowledge. So, yes, started in the storage subsystem, but absolutely where we're heading is a memory-driven computing world where it's in the memory subsystem. And that is the machine which is a big R&D project coming out of HP Labs and Martin Fink showing it off a couple of years ago and giving us the roadmap and now we're seeing it evolve. But as well, I would think your ecosystem can kind of build its own pseudo machine with compute power and all this persistent memory and architecting flat. I mean, do you see those as two separate vectors that let's see what happens in the channel or do you see those two worlds coming together? Fundamentally, we are a partnering business, right? We work extensively with the ecosystem of software providers, partners in the industry. So what we're demonstrating with the machine, necessarily we've done a lot of those parts ourselves to show the capability, but absolutely this memory-centric computing vision and that we're beginning to realize with some of the products that we're releasing today, persistent memory is a great example of that, is all about enabling the industry, the ecosystem and the industry to bring that value for ultimately for all of our customers. And as heading up the sort of options business, if you will, how do you, one of the things we've talked about a lot is this notion of true private cloud, which is substantially mimicking public cloud on-prem because the world is hybrid, as we all sort of point out. How do you create that experience for customers, that cloud-like experience? I think well with the simplicity and the agility of what we're doing with the HPE compute experience now is very focused on creating that cloud-like experience in a hybrid cloud world, right? On-premises for example. And so that's, a lot of that is about being able to scale up and down the computing capability to incorporate new financial models so that you can buy compute, rent compute at your kind of, depending on what your strategic corporate objectives are. So the options themselves then give you that ability for example, scalable persistent memory uses the base system in the server. And one day you may say, hey I need to use a portion of that as persistent because of the workload that I'm running and then later that night, that same system could translate over to run a completely different workload and change the profile of the persistent memory that's being used because it's a configuration setting of the base system memory. So making the system itself very flexible to adapt to the changing needs of workloads either over time or very real-time like that in the course of a day. Tom, I want to get your thoughts on the trends that we're covering and certainly the industry's covering so you have an industry scope and you have, I'll see partners which are going to be critical in getting things certified and or working so the customer just plugs stuff in like the memory. Obviously the marketplace starts, oh, server shipments are down, but the cloud's happening, servers are actually growing, depending on how you look at it. But there's more real-time stuff going on, there's more processing happening. How do you guys look at the marketplace trends because there is more need at the edge, for instance. We had Bill, Bill been on just earlier before you came on talking about how storage and computer coming together. This is kind of the options world you're in, you're in the middle of all this actually. So people actually coupling together and composing solutions, whether it's on-prem or working with similar architectures in the cloud, same code bases moving back and forth. So this whole world is really not declining, maybe shifting how it was before that transformation you're in the middle of. How do you guys look at that and how do you talk to customers who are like, I was buying servers and options before, I still got to do that but I got to transform and be prepared for real-time analytics, using multiple clouds, all these kinds of these important items for the future. It's a bit of an architectural revolution if you will, right? As we move to a memory-centric computing world, as we move to a world where everything is cloud-based, cloud architecture-based, whether it's out in a public cloud somewhere on an on-prem is kind of hybrid IT model, really is a completely different architectural model. And so to capitalize on that, what we work with customers on is things like the composable capability of synergy, things like persistent memory and making that scalable to move to a memory-driven computing model. Things like our all-flash array business and our product offerings to be able to accelerate that storage subsystem well beyond what's been done. It's performance-driven too. You've got to eke out performance more and more. That's kind of the mandate. Another interesting thing I want to get your thoughts on and I'm old in the industry these days relative to the average age of most people in the big companies like hyperscalers or like 28 to 30 something. The trend is systems. I mean, you're seeing, if you look at what the cloud's doing and the revolution at the architectural level, it's almost a complete crossover to a systems mindset. Systems meaning operating systems or core systems. So a lot of the people that are really doing well in the industry who are radically transforming it are older guys. James Gosling was at, just joined Amazon Web Services. You've got guys who are in their 50s who are leading major architectural shifts. This kind of puts HP in an interesting position because you guys have so much experience with systems, servers just on an isolated basis. But now as that looks out over the landscape, it's even more important to look at it from a holistic perspective. Your thoughts on this trend? Yeah, well, absolutely. It's a huge trend. And by taking the expertise that we've got at a systems level and coupling that with our strategic imperative to partner with other industry leaders in the industry revolution, I think those two together position, not just HPE, well, but the ecosystem of HPE and our software industry partners to really help advance that architectural. James, governor last night who's with Red Monk and one of the research friends we like and I said to him, it's open bar and open source. It was my kind of headline story I was trying to, we were collaborating on. And what I mean by that is that, as open source evolves, we saw some of the stuff going on with the machine in Memster, a lot of that stuff's going to be open source at the core, at the system level. So open source is growing, but when I say open bar, it's like there is more goodness being contributed to open source than ever before. You're seeing great machine learning libraries being given in to open source for collaboration. You're seeing open source being a great recruiting environment. So if you're a young gun in the marketplace right now and you're getting all this contribution, so with that kind of as a context, what's the open source strategy that you see because you're in kind of a glue layer with options. You're kind of creating some flexibility for customers. At the same time, you've got a glue kind of concept going on with software. What do you guys do with open source? Is there a trend there that you'd like to share? I mean, I'm interested to know what your position is, vis-a-vis contribution programs and whatnot. Yeah, I think certainly at the hardware layers, right, of what we're doing with server options, it's about enabling new capabilities. And so we work with quite a few open source partners to enable those. Suse, for example, Red Hat, they're taking our persistent memory offerings and optimizing those so that they get certainly the immediate benefit of a layer of high-performance storage, but the more radical performance improvements that they can get when they address directly a layer of persistent memory. So it's not so much that we're creating a whole new, at least in the option space, kind of a whole new open-source play. But you're intersecting. Absolutely. I mean, one of the things, networking is hot right now, certainly SDM, we see that. A lot of the open-source projects, even in the Linux foundation, you're seeing the network stack just kind of being kind of decomposed. So a very interesting opportunity. Yes. It was storage, right? And you mentioned Flash a couple of times, right? You're seeing the whole storage stack just completely morphing and changing. So you guys are in the center of that. How does a customer engage? Does it happen typically through the channel? Do they go to HPE.com? How does that happen? For server option kind of products, yeah. Certainly, through our direct sales force, through our partner's sales forces, because as we said, it's an ecosystem that brings this value forward. So in many cases, it's not us even in explaining something like persistent memory. It's Microsoft, or it's Red Hat, or it's SUSE, or other partners, they're VMware. Got actually a lot of presence with VMware here and some interesting things they're going to be talking about in one of our sessions here later today at Discover. So that's one path, or two paths, our sales force, their sales force, and then absolutely the channel. We've got a very rich channel program and a lot of engagement with them to bring them up on networking technology, storage technologies, memory, persistent memory technologies, so that they can effectively engage customers, yeah. All right, Tom Latin is the Vice President and General Manager of HPE Server Options. My final question for you to end the segment here is, what should customers know about Gen10 and server options if you had a chance to look right in the camera and say, hey, new game in town or think differently around architecture, what would be your words and your words? What should customers know about the world you're building? Certainly with Gen10, the Server Options portfolio unleashes or helps support the overall security capabilities of Gen10, number one, but if I can have a second one. Of course. I've got to play persistent memory high because we've got now a terabyte scale persistent memory capability in the Gen10 platforms, which opens up a whole new world of opportunity for applications, as I said early on, to develop or increase performance, in many cases, 20, 25, 27 times the capabilities today. That's awesome. I mean, I think the memory is awesome. Dave and I have been talking for years that memory used to be the resource that was constrained and unlimited storage. Now it's the other way around. People want memory, application developers and programmers. It's a cube bringing you great content you can put to memory, flash memory, HPE Discover, I'm John Furrier, Dave Vellante. We'll be back with more live coverage after this short break.